Pidgin 3 - Is there any point?

[Gary Kramlich]

On Thu, Jan 24, 2019 at 12:42 AM Eion Robb <eion at robbmob.com> wrote:

> Hi all,
>

Sorry for the late response, I need to get better notifications for these
emails or something..


> Forgive the clickbaity subject line, but I'm hoping to open some
> discussions.
>

Discussion is always good ;)


> Over the past few years, there's been a few times where I've though "oh,
> it'd be good to send a PR to fix up XYZ issue with Pidgin2" but then my
> response is "well, I shouldn't do that because everything is going on in
> Pidgin3" and then "well, Pidgin3 is a moving target that I'm not personally
> using so I can't test my changes, so I better not break things for other
> people, so I won't make a PR"
>
> So this got me thinking.....
>
> Is there any real use for users to adopt Pidgin 3?  Is there any "killer
> feature" that means they should want to give up on Pidgin 2 and get
> mass-adoption of the new and shiny, or are we just making work for the sake
> of work, and breaking everything in the process?
>

Right now, no we are not where I hoped to be.  There's lots of reasons for
that, but the biggest is that we've been stuck in churn hell trying to
remove webkit1gtk which has now been removed from arch, fedora, and debian
testing.  The killer features I've been chasing are plugins in other
languages (ie gplugin), voice and video, saner security defaults with the
less secure options being opt-in (ie logs off by default, forced password
store, not clear text file), and other stuff along those lines.  You can
see more on my personal pidgin trello board here
https://trello.com/b/4ZBlhJFd/pidgin


> Because right now, the way I see it from a user perspective is:
>
> *Pros*:    Gtk3 (does this even matter?), possible pro: "v/v on windows"
> *Cons*:   All my plugins/themes/customisations don't work
>

To most users, gtk3 doesn't matter... unless you're on osx where pidgin3 no
longer requires xquartz.  I assume with the modular gdk backends that
wingdi is better too, but I haven't tested it personally, so I have no idea.

So I want to ask everyone else's opinions....
>
> Have we shot ourselves in the foot, trying to clean up the code without
> any actual end-user-benefit?
>

I think we dumped way too much into pidgin3 a long time ago and we're still
trying to unbury ourselves from it.  Case in point..  iirc last time I
looked up when we "announced" the 3.0 branch, the commit date was in 2011
(but maybe that was the hg conversion?) but we have a GSoC project from
2009 that's waiting on 3.0.0.


> Is Pidgin3 the next "PHP6"?
>

I don't know the story of php6, so no?


> Are we better off focusing our efforts on Pidgin 2 and merge in all the
> plugins/hacks/workarounds that people have made, to improve it?
>

I have no intention on doing this.  We have made some serious gains on
pidgin3/purple3 when it comes to code/api cleanups and I don't want to do
that again.  I use pidgin 3 every day, sure there are bugs, but it's not
crashing and it runs fine for me.


> Or, do we need to release a Pidgin2 plugin loader that can load the "100
> most popular plugins" for Pidgin3 to be usable and have a safe transition
> process?
>

I believe we've talked about this before, and this concept was the entire
point of https://bitbucket.org/pidgin/twotree.  Also I've been trying like
hell to get a plugin site built that we can check for this stuff, either
via build agents or what have you, but I haven't even been able to go
through the plugin pack yet.. so yeah, I personally won't be tackling
that.  But the idea of the plugin site is to make
developing/finding/installing/updating plugins easier for everyone.  I can
talk more on this, if anyone would like me to.


> Thoughts, comments, criticisms? :)
>

I have to be 100% honest here.. When I first stepped up to take over the
project I told a few of you that I probably want to do this for about 2
years as I don't/can't work a full time job and be the maintainer pidgin
needs for much longer than that.  And guess what, I was right.  This has
become exceedingly difficult the past few months right as we're hitting the
2 year mark.  Basically, I'm burning out hardcore...

But that said, we're finally starting to see that light... We're nearly
ready to make pidgin3 installable side-by-side with pidgin2 which is
necessary for all of the points you mention about plugins above.  We need
people to be able to switch back and try again later as we stabilize the
api.  Then again this wouldn't necessarily have been necessary if we split
the repos out earlier as we could have stabilized the libpurple api and
released that separately for bitlbee, spectrum2, chatty, sapphire, adium,
etc to get a feel for while we then focused on getting pidgin2 there.  But
none of that is an option anymore because we have to do it all at once for
a myriad of reasons like we have to finish porting off of webkit because
it's being removed from distributions.

I'm realizing I'm kind of getting ranty, which is for a number of reasons,
but lets just say it's because I'm tired.

Regardless I'd love to discuss this more, but doing so via email is going
to get extremely tedious extremely fast.  If you would perhaps like to do a
voice call, or just talk while streaming on twitch (for recording's sake
and including others), that is something we can 100% do.


> Cheers,
> Eion
>

Thanks,

--
Gary Kramlich <grim at reaperworld.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pidgin.im/pipermail/devel/attachments/20190131/ba41b158/attachment.html>